Chemical reactions such as reductive alkylation of azides (J. Org. Chem. 1995, 60, p 4986-4987) or reductive ring-opening reaction of oxazolidine derivatives (Tetrahedron Letter 39 (1998) 1985-1986) and the like have been known for the production of N-substituted amino acids.
Although microbiological methods for producing amino acids from 2-oxocarboxylic acid derivatives using a dehydrogenase or aminotransferase are known, these methods mainly include simple amination. The known methods using a substituted amino group are only a simple methylamination method such as the method of producing N-methylalanine from pyruvic acid using microorganisms of genus Pseudomonas (J. Biol. Chem., 250, p 3746-3751 (1975)) and the methylamination method using microorganisms of genus Rhodococcus and Arthrobacter (JP Patent Publication (Kokai) No. 2001-190298).
With respect to enzymes involved in the above reaction, it is reported in J. Biol. Chem., 250, p 3746-3751 (1975) that N-methylalanine dehydrogenase from Pseudomonas MS ATCC 25262 was purified.
On the other hand, as for a method of producing cyclic amino acid chemically, such methods have been known as producing L-azetidine-2-carboxylic acid (Stephen Hanessian et al., Bioorganic & Medicinal Chemistry Letters (1999) vol. 9, pp. 1437-1442, and U.S. Pat. No. 5,942,630); pipecolic acid (Concepcion F Garcia et al., Tetrahydron Asymmetry (1995) vol. 6, pp. 2905-2906); 4- and 5-hydroxypipecolic acid (Roland E. A. Callens, et al, Bull. Soc. Chim. Belg. vol. 91, (1982) pp 713-723); 1,4-thiazane-3-carboxylic acid (Biosci. Biotechnol. Biochem., vol. 62, pp 2382-2387 T Shiraiwa, et al.) (Acta Chemica Scandinavica, 1994, vol. 48, pp 517-525, U Larsson et al.); L-3-morpholine carboxylic acid (Bull. Chem. Soc. Jpn., vol. 60, pp 2963-2965, 1987, Y Kogami et al.); (S)-azepane-2-carboxylic acid (Liebigs. Ann. Chem. 1989, pp 1215-1232, D. Seebach et al.); and the like.
As for a method of producing cyclic amino acid biochemically, such methods have been known as producing L-pipecolic acid from L-lysine while utilizing pyrroline-5-carboxylate reductase (EC 1.5.1.2) (Tadashi Fujii et al., Bioscience Biotechnology Biochem (2002) vol. 66, pp. 1981-1984); L-proline from L-ornithine while utilizing pyrroline-5-carboxylate reductase (EC 1.5.1.2) (Janet Kenklies et al., Microbiology (1999), vol. 145, pp. 819-826; and Ralph N Costilow et al., Journal of Bacteriology (1969) vol. 100, pp. 662); L-proline from L-ornithine with ornithine cyclodeaminase (Ralph N Costilow et al., Journal of Biological Chemistry (1971) vol. 246, pp. 6655-6660); various types of cyclic amino acids from various types of diamino acids with ornithine cyclodeaminase (International Publication WO 02/101003); and the like.
On the other hand, as for an enzyme that reduces a cyclic amino acid having a double bond at 1-site, pyrroline-2-carboxylate reductase: EC1.5.1.1, for example, derived from animal or fungus is known as the enzyme that reduces Δ-1-pyrroline-2-carboxylic acid and Δ-1-piperidine-2-carboxylic acid to generate proline and pipecolic acid respectively (Alton Meister et al., Journal of Biological Chemistry (1957) vol. 229, pp. 789-800).
Further, there is a report that describes such metabolism of a bacterium belonging to Pseudomonas species as generating L-pipecolic acid from D-lysine through Δ-1-piperidine-2-carboxylic acid as an intermediate, and that piperideine-2-carboxylate reductase: EC 1.5.1.21 conducts the reduction reaction in the reactions (Cecil W Payton et al., Journal of Bacteriology (1982) vol. 149, pp. 864-871).
In addition, it has been found that ketimine-reducing enzyme: EC 1.5.1.25 derived from liver of porcine reduces S-aminoethylcysteine ketimine, lanthionine ketimine and cystathionine ketimine (Mirella Nardini et al., European Journal of Biochemistry (1988) vol. 173, pp. 689-694).
However, the method reported by Fujii et al. (Tadashi Fujii et al., Bioscience Biotechnology Biochem (2002) vol. 66, pp. 1981-1984) includes steps of using L-lysine 6-aminotransferase for L-lysine to generate Δ-1-piperidine-6-carboxylic acid as the intermediate, and further acting the reductase on it to give L-pipecolic acid. The method can deal with only the case where the starting material is L-lysine, and can not be applied to the production of other cyclic amino acids.
The report of Costilow et al. (Ralph N Costilow et al., Journal of Biological Chemistry (1971) vol. 246, pp. 6655-6660) describes the step of obtaining L-proline by using Ornithine Cyclase for L-ornithine, but also does not describe any products other than proline. Denis et al. (WO 02/101003) disclose a method of obtaining L-pipecolic acid, L-Thiomorpholine-2-carboxylic acid, 5-hydroxy-L-pipecolic acid and the like by using Ornithine Cyclase, but do not describe yield, optical purity and the like.
In any of the aforementioned methods, the optical purity of the produced cyclic amino acid depends on the optical purity of a starting amino acid, and an optically active cyclic amino acid can not be obtained from a starting material of whole racemic body with a high yield.
On the other hand, a method employing a cyclic amino acid having a double bond at 1-site as an intermediate is advantageous industrially, because it can use racemic cyclic amino acids or diamino acids.
Enzyme reactions that deal with L-proline and L-pipecolic acid, L-Thiomorphine and the like, respectively, are confirmed. However, in each case, an enzyme reaction is confirmed only biochemically, and no example of industrial production has been known. Further, there is such a description that enzymes derived from animal are very unstable, making practical application difficult by using these enzymes.